GRUAN ICM-11, Singapore, May 23, 2019
Takuji Sugidachi, (Meisei Electric Co. LTD., Japan)
Meisei SKYDEW instrument: analysis of results from Lindenberg campaign
1
Overview The test comparison campaign was conducted at Lindenberg, on April 2019 1) Chilled-mirror hygrometers SKYDEW and CFH 4 soundings 2) Cloud particle sensors COBALD and CPS 1 sounding 3) Operational radiosondes
Meisei iMS-100 and Vaisala RS41 12 soundings
CPS COBALD
CFH SKYDEW
RS41 iMS-100 2
Output data to Meisei Radiosonde or Other radiosonde (XDATA)
Peltier
Optical sensor
PT100
Peltier-based chilled-mirror hygrometer “SKYDEW”
Features of SKYDEW hygrometer 1. Two-stage Peltier device additional cooling system with the latent heat of evaporation for liquid ethanol No need to use cryogen material (CHF3) 2. Dew/frost detection by scattered light using an electronically modulated light 3. Digital controller (PID controller, gain
scheduling depending on dewpoint)
4. Meisei original data format or XDATA format
Sensor
Heatsink
Control board
Battery
SKYDEW has been developed since 2009 by Meisei and Hokkaido university.
Ethanol
Optical fiber
Ice Mirror
Heat Sink
3
Date Place SN Comments
FL10 2018.01.29 17:30(LT) Moriya , Japan 18002
FL11 2018.02.27 17:30(LT) Moriya , Japan 18003
FL12 2018.04.20 14:45(LT) Tateno , Japan 18001 with CFH by JMA*
FL13 2018.07.15 16:46(LT) Kototabang, Indonesia 18005 with CFH by JAMSTEC*
FL14 2018.07.16 17:00(LT) Kototabang, Indonesia 18006 with CFH by JAMSTEC
FL15 2018.07.19 20:40(LT) Kototabang, Indonesia 18007 with CFH by JAMSTEC
FL16 2019.02.15 14:00(LT) Moriya , Japan 18011
FL17 2019.03.22 14:30(LT) Tateno, Japan 18013 with CFH by JMA
FL18 2019.04.01 21:23(LT) Lindenberg, Germany 18012 with CFH by GRUAN LC*
FL19 2019.04.02 14:30(LT) Lindenberg, Germany 18010 with CFH by GRUAN LC
FL20 2019.04.04 11:46(LT) Lindenberg, Germany 18999 with CFH by GRUAN LC
FL21 2019.04.04 11:46(LT) Lindenberg, Germany 18015 with CFH by GRUAN LC FL22 2019.04.05 12:31(LT) Lindenberg, Germany 18010 with CFH by GRUAN LC
Test flights & Comparison with CFH 13 soundings of SKYDEW(product type) have been conducted since 2018. The comparisons with CFH are10 times at Tateno, Lindenberg, and Kototabang. 1. Performance of SKYDEW (e.g., cooling limits) 2. Comparison results
*The sounding data of CFH are provided by JMA, JAMSTEC, and GRUAN Lead center, respectively 4
Performance at nighttime (e.g., 1 April, 2019, Lindenberg)
Mirror-heating for refresh of ice on mirror.
Cooling limit at 25km
FP/DP
Heatsink
The Scattered light corresponds the size of ice/water on the mirror. --> It should be constant.
Larger Peltier-current can make larger temperature gap between mirror and heatsink.
5
Performance at daytime (e.g., 2018.07.16, Kototabang)
Cooling limit at 25km
FP/DP
6
Performance at daytime (e.g., 2019.04.05, Lindenberg)
Cooling limit at 16.5 km
FP/DP
7
Cooling limit
Performance at daytime (dependence on latitude)
@Lindenberg (52N 14E) @Tateno (35N 140E) @Kototabang(0S,100.3E)
Cooling limit
Cooling limit
The measurement range is restricted by the performance of Peltier cooling. 1. Daytime performance is limited as compared with nighttime performance, because the heatsink (hot side of Peltier) is heated by solar radiation. 2. The measurement at higher latitude is more difficult than that at tropics, because tropopause is lower and need to make temperature gap for longer time.
8
White-painted heatsink (This picture show black type)
Improvement of heatsink cooling
9
Comparison between SKYDEW and CFH
SKYDEW
CFH
The comparison with CFH is total 10 sounding since 2018. The mirror temperature of SKYDEW and CFH were compared at each 100m.
Cooling limit for SKYDEW
SKYDEW - CFH
10
Comparison between SKYDEW and CFH
Date Place
FL12 2018.04.20 14:45(LT) Tateno
FL13 2018.07.15 16:46(LT) Kototabang
FL14 2018.07.16 17:00(LT) Kototabang
FL15 2018.07.19 20:40(LT) Kototabang
FL17 2019.03.22 14:30(LT) Tateno
FL18 2019.04.01 21:23(LT) Lindenberg
FL19 2019.04.02 14:30(LT) Lindenberg
FL20 2019.04.04 11:46(LT) Lindenberg
FL21 2019.04.04 11:46(LT) Lindenberg
FL22 2019.04.05 12:31(LT) Lindenberg
The temperature difference of total 8 sounding are plotted. For troposphere, the difference is less than 0.3 K expect two soundings(FL19, FL22) *There are possibility that the calibration of the SKYDEW used for FL19 and FL22 (SN18010) had small bias because of the calibration method. For stratosphere, the difference is over than 0.5 K. The bias of FL20 is positive, while the others are slightly negative.
SKYDEW - CFH
FL20
11
(18012)
Comparison between SKYDEW and CFH Same SKYDEW was used for the sounding of FL18 and FL20.
(18012)
--> Regarding with this bias (FL20), further investigation is needed.
FL18 Nighttime
FL20 Daytime
Raw CFH Averaged CFH
Oscillation of CFH at noon
12
The payload of FL18 was recovered, and reused for the sounding of FL20.This means that this difference is not derived from the calibration error. The CFH profile of FL 20 have large oscillation, which may be related to the solar contamination. However, it is considered that the oscillation does not generate bias.
Comparison between SKYDEW and CFH
Date Place
FL12 2018.04.20 14:45(LT) Tateno
FL13 2018.07.15 16:46(LT) Kototabang
FL14 2018.07.16 17:00(LT) Kototabang
FL15 2018.07.19 20:40(LT) Kototabang
FL17 2019.03.22 14:30(LT) Tateno
FL18 2019.04.01 21:23(LT) Lindenberg
FL19 2019.04.02 14:30(LT) Lindenberg
FL20 2019.04.04 11:46(LT) Lindenberg
FL21 2019.04.04 11:46(LT) Lindenberg
FL22 2019.04.05 12:31(LT) Lindenberg
Excluding the doubtful sounding (contamination at cloudy condition) and the sounding of FL20, the difference temperature between SKYDEW and CFH is less than 0.5 K roughly. At stratosphere, it seems that the SKYDEW are slightly negative in comparison with CFH. --> Currently, the cause of this bias is under investigation.
SKYDEW - CFH
13
(18012)
Comparison between SKYDEW and CFH
One of the possibility to derive this bias is the oscillation by the PID controller. The aggressive PID setting causes the larger oscillation. The oscillation can be smoothed out using appropriate filters (Vömel et al, 2016) The oscillation may be too large for the SKYDEW, although the presence of slight oscillation is preferred. If the center of the oscillation is not true frostpoint, this smoothing of oscillation cause the bias.
Averaged mirror temp(SKYDEW) Raw mirror temp(SKYDEW)
CFH Ascent
CFH Descent
14
Bias caused by PID oscillation ? The experiment was conducted to investigate the negative bias. Two SKYDEWs measure same air, and the PID parameter for one SKYDEW was changed to make oscillation for a minute. When the oscillation is large, it seems that the averaged mirror temperature indicates lower than the other SKYDEW.
SKYDEW 1
SKYDEW 2
Aggressive PID Aggressive PID
Raw (1sec)
Smoothed (+/- 6sec)
Is this systematic bias or random error? --> Further investigation is needed.
15
Smoothed (+/- 6sec) Raw (1sec)
Dual sounding of SKYDEW
At 4 April 2019, two SKYDEW instruments were flown on the same payload to evaluate the reproducibility. The difference between the measured frost point temperatures is almost less than +/-0.2 K up to 16.5km (cooling limit)
Cooling limit Cooling limit
SKYDEW1
SKYDEW2
16
Summary
1. SKYDEW can measure dewpoint/frostpoint up to 25 km at nighttime, and up to 15-25 km at daytime depending on locations.
2. The comparison with CFH shows good agreement at troposphere. The SKYDEW has slightly bias as compared with CFH at the stratosphere. The cause of this bias is still under investigation.
3. The dual sounding with two SKYDEW shows good agreement. The difference between the measured frost point temperatures is almost less than +/-0.2K.
4. The product model of SKYDEW will be released on October 2019.
We thank JMA (Japan metrological agency) for providing the CFH data at Tateno, Dr. J. Suzuki at JAMSTEC (Japan Agency for Marine-Earth Science and Technology) for providing the CFH data at Kototabang, and the GRUAN lead center and the staffs at the Lindenberg observatory for the support of the observations at Lindenberg 17
Thank you for your kind attention.
For more information, contact “[email protected]”
18
19
Comparison flight with CFH
Temperature Relative humidity +CFH
Three RH profiles have good agreement basically. When changing RH, the difference becomes large due to the response time or the tuning of TL correction. (RS41 > CFH > iMS-100)
The iMS-100 and RS41 are flown with the CFH.
20
Comparison flight with cloud particle sensors The iMS-100 and RS41 are flown with the particle sensors, COBALD and CPS(Cloud particle sensor) produced by Meisei.
Temperature Relative humidity Cloud particle
Layer 1 (no cloud layer): The RH profile shows high supersaturation. Layer 3,4 (could layer): The RH profile shows slightly supersaturation.
Cloud layer
No Cloud
No Cloud
Layer 1
Layer 2
Layer 3
Layer 4
The information about existence of cloud is useful for the evaluation of RH value. 21
Result of CPS Ascent T, RH, U/V, Particle Number, Particle size, Particle shape
Number concentration from number of count per second (>2 μm)
I55 , Scattered light level ---> particle size (2um – 140um)
The degree of polarization (DOP) ---> particle shape
Correction value (Fujiwara et al. 2016, AMT)
22
気圧を段々と下げていく
Ethanol(15g) cool hotside 15℃ Condition peltier current: 1.2A (constant) Pressure :1000hPa ⇒30hPa(during 1200sec)
1000 500 250 125 60 30
Pressure is decreased gradually.
Peltier current ON (constant 1.2A)
Ethanol cooling start
●innovation of hotside cooling A chilled-mirror hygrometer by peltier cooling is needed to cool hotside effectively. But, air cooling are not expected because of a little air in upper air, in particular stratosphere. This sensor use cooling by heat of evaporation. When ethanol evaporate, it take heat from air. In this case, ethanol take heat from heatsink.
Heat of evaporation(ethanol) : 853J/g Heat capacity of hotside : about 30 J/K
Ethanol is used Ethanol is not used
1000hPa
500hPa
200hPa
23
Cooling capability by Peltier device
We estimate that ΔT>30@LS, ΔT>55@surface.
Dependence on (a)air temperature, (b)pressure, (c) airflow
Aggressive cooling of the hot side In addition to Peltier cooling , we use liquid ethanol for cooling of the heatsink. When ethanol evaporates on surface of the heatsink, the latent heat of evaporation can provide additional cooling effect of heatsink.
Cold side
Hot side
heatsink
PAT. P.
24
(T=-21℃、-31℃FP)
(T=-39℃、-45℃FP)
(T=-60℃、-61℃FP)
Ice size: 10 – 20 um
5 – 10 um
2 – 5 um
100 um
100 um
100 um
R= 4 um ISL =80 mW 400 個/mm2 32 W
R= 2 um ISL =10 mW 2880個/mm2 28.8 W
Ice on the mirror
The ice size and density on the mirror depends on the mirror temperature. > large ice and low density for high T small ice and high density for low T The growth/evaporation rate of larger ice is lower, which cause slower response particularly in UT/LS. Ice on mirror need to be refreshed by heating mirror in cold enough conditions in a similar to the CFH.
2 mm
25
Comparison between SKYDEW and CFH @ Lindenberg, 20190401-05 SKYDEW and CFH have a good agreement (+/- 0.5degC) at troposphere. The measurement biases are over than 0.5degC at stratosphere.
(18012) (18010)
(18010)
26
Comparison between SKYDEW and CFH @ Kototabang
27
2019.4.1 2019.4.2 2019.4.4 2019.4.5
Water vapor mixing ratio of SKYDEW, CFH, and AURA MLS
28
29
Meisei SKYDEW instrument: �analysis of results from Lindenberg campaign�Foliennummer 2 Peltier-based chilled-mirror hygrometer “SKYDEW”Test flights & Comparison with CFH Foliennummer 5Foliennummer 6Foliennummer 7Foliennummer 8Foliennummer 9Foliennummer 10Foliennummer 11Foliennummer 12Foliennummer 13Foliennummer 14Foliennummer 15Foliennummer 16Foliennummer 17Foliennummer 18Foliennummer 19Foliennummer 20Foliennummer 21Foliennummer 22Foliennummer 23Cooling capability by Peltier device Ice on the mirror Foliennummer 26Foliennummer 27Foliennummer 28Foliennummer 29
Top Related